Processing and Analytics Vasiliki (Vasia) Kalavri
 vkalavri@bu.edu Spring 2020 4/09: Flow control and load shedding ??? Vasiliki Kalavri | Boston University 2020 Keeping up with the producers • Producers (back-pressure, flow control) 2 ??? Vasiliki Kalavri | Boston University 2020 Load management approaches 3 ! Load shedder (a) Load shedding (b) Back-pressure (c) Elasticity Selectively drop records: Suitable for transient load increase. Scale resource allocation: • Addresses the case of increased load and additionally ensures no resources are left idle when the input load decreases. ??? Vasiliki

Discovery ................................................................................ 6 1.3.7 Load Balancing............................................................................................. an Application .....................................................26 3.3 Load Balancing using Rancher Load Balancing services ............................................27 3.4 Service Discovery needs robust cluster management capabilities that can handle scheduling, service discovery, load balancing, resource monitoring and isolation, and more. For years, Google has used a cluster manager

End of 2021 100% services migrated to Istio 8 Features currently used: ● HTTP/2 Load-balancing ● Traffic Shifting ● mTLS Features under investigation: ● Retries ● Circuit breaking Istio Istio ● Moving HTTP/2 load-balancing from client-side to Envoy ● Label selector updates for app and version labels ● Istio default retry policy ● Istio proxy performance and load testing ● Abstracting features 44 Moving HTTP/2 load-balancing from client-side to Envoy Adopting Istio ● We use gRPC heavily in our microservices ● But Kubernetes is pretty bad at load-balancing it ● So we solved it by

selectivity • Dataflow optimizations • plan translation alternatives • Runtime optimizations • load management, scheduling, state management • Optimization semantics, correctness, profitability Topics University 2020 34 • Fission might be preferable to pipeline and task parallelism because it balances load more evenly • Data-parallel streaming languages enable fission by construction • Elastic scaling qualified: if load balancing is applied after fission, each instance must be capable of processing each item and have access to necessary state • Establish placement safety: if load balancing while performing

& Authorization CHAPTER SIX HIGH AVAILABILITY Kyuubi provides both high availability and load balancing solutions based on Zookeeper. 6.1 Quick Start In this section, you will learn how to setup copy of hive-site.xml into $$SPARK_HOME/conf, every single Spark application will automati- cally load this config file to its classpath. This version of configuration has lower priority than those in without requiring human intervention. 6.2. Deploying Kyuubi 55 Kyuubi, Release 1.3.0 Load Balance Mode Load balancing aims to optimize all Kyuubi service units usage, maximize throughput, minimize response

& Authorization CHAPTER SIX HIGH AVAILABILITY Kyuubi provides both high availability and load balancing solutions based on Zookeeper. 6.1 Quick Start In this section, you will learn how to setup copy of hive-site.xml into $$SPARK_HOME/conf, every single Spark application will automati- cally load this config file to its classpath. This version of configuration has lower priority than those in without requiring human intervention. 6.2. Deploying Kyuubi 55 Kyuubi, Release 1.3.0 Load Balance Mode Load balancing aims to optimize all Kyuubi service units usage, maximize throughput, minimize response

shell:stack-traces-print shell:threads shell:while log:list bundle:capabilities bundle:diag bundle:id bundle:load-test bundle:requirements bundle:resolve system:name We encourage the users to use the --help option KARAF.plist sudo chmod u=rw,g=r,o=r /Library/LaunchDaemons/org.apache.karaf.KARAF.plist launchctl load ~/Library/LaunchAgents/org.apache.karaf.KARAF.plist launchctl start org.apache.karaf.KARAF launchctl At runtime, when you change the SSHd server configuration, you have to restart the SSHd server to load the changes. You can do it with: karaf@root()> bundle:restart -f org.apache.karaf.shell.ssh The

period. HA works to provide Kyuubi that meets an agreed-upon operational performance level. • Load balancing – It becomes necessary for Kyuubi in a real-world production environment to ensure high availability achieve zero downtime for planned system maintenance • Failure detectability – Failures and system load of kyuubi server and engines are visible via metrics, logs, and so forth. 4 Chapter 1. A Unified and Flink to build a streaming data warehouse. And then, you can use Zookeeper to enable the load balancing for high availability. The data could be stored in Hive, Apache Iceberg, or other DBMSs. In

period. HA works to provide Kyuubi that meets an agreed-upon operational performance level. • Load balancing – It becomes necessary for Kyuubi in a real-world production environment to ensure high availability achieve zero downtime for planned system maintenance • Failure detectability – Failures and system load of kyuubi server and engines are visible via metrics, logs, and so forth. 4 Chapter 1. A Unified and Flink to build a streaming data warehouse. And then, you can use Zookeeper to enable the load balancing for high availability. The data could be stored in Hive, Apache Iceberg, or other DBMSs. In

period. HA works to provide Kyuubi that meets an agreed-upon operational performance level. • Load balancing – It becomes necessary for Kyuubi in a real-world production environment to ensure high availability achieve zero downtime for planned system maintenance • Failure detectability – Failures and system load of kyuubi server and engines are visible via metrics, logs, and so forth. 4 Chapter 1. A Unified and Flink to build a streaming data warehouse. And then, you can use Zookeeper to enable the load balancing for high availability. The data could be stored in Hive, Apache Iceberg, or other DBMSs. In